First define the claim you are testing
“Only noticing what is going wrong” is partly an attention problem and partly an evidence problem. The belief behind it is often something like “The problems are the only honest part of the picture.” To challenge that belief, you do not need to prove that everything is good. You need evidence showing that the negative scan is incomplete.
Use a three-column daily record
For seven days, write three short entries at the end of the day: problem, neutral fact, and thing that worked. The categories matter. A neutral fact might be “the meeting happened on time.” A thing that worked might be “I sent the email I had been avoiding.” Neither has to feel exciting.
If the negative column is full and the other two remain empty despite careful looking, that supports the original belief. If all three columns fill repeatedly, the belief that problems are the only honest evidence becomes too broad.
Look for prediction error
Before a difficult task, write what you expect to go wrong. Afterwards, compare the forecast with what happened. Did every feared problem occur? Were some problems smaller, later or different from expected? Did anything go normally?
Prediction error is powerful evidence because it compares a belief with a specific expectation rather than with a vague mood.
Count what your attention skipped
Review a day that felt completely bad and list events in chronological order. Notice how many ordinary actions disappeared from memory: meals eaten, messages answered, tasks completed, transport that worked, people who behaved normally. These are not “gratitude wins.” They are missing data.
The point is not to celebrate everything. It is to see whether your first summary compressed a mixed day into a single negative label.
Challenge the belief with behaviour
Choose one small action that usually feels pointless when the negative scan is active. Complete it, then record the outcome without grading the whole day. If the action goes reasonably, that does not prove life is good. It does show that “nothing is working” was not a complete description of that hour.
Use outside evidence carefully
Ask someone you trust for one concrete observation about the situation, not reassurance. “What did you notice about how that meeting went?” is more useful than “Do you think I did badly?” Compare their observable details with your own first summary.
Different perspectives can reveal selective attention, but they should not be used to dismiss genuine problems.
Keep real problems visible
Balanced evidence is not a command to minimise harm, debt, conflict, illness, discrimination or other serious difficulties. If something is wrong, it belongs in the record and may need direct action. The challenge is to stop one valid problem from automatically becoming the only fact allowed to count.
Define what would change your mind
Write a rule in advance. For example: “If I can identify at least one neutral or working element on five of seven days, I will replace ‘everything is going wrong’ with ‘some important things are going wrong, and some parts are still functioning.’”
That is a narrower, more defensible statement. It does not require positivity; it requires proportion.
Use the belief-gap tool as a sample tracker
In the Self-Concept Audit / Belief Gap Finder, save the original summary, the three-column entries and the revised wording. Review after a week rather than after each mood swing. The useful evidence is the pattern across days.
One more check before closing the diagnostic
For What Evidence Would Challenge Only Noticing What Is Going Wrong?, write one observable sign that would show the current recommendation is helping and one sign that would show it is not. Review those signs at the next planned check rather than changing the approach in the middle of the experiment. This gives the user a clear exit condition and keeps Pulse from turning uncertainty into endless advice.
Bottom line
Evidence that challenges only noticing what is going wrong is not forced good news. It is a wider sample: problems, neutral facts, working elements and prediction errors. If that broader record repeatedly shows more than the negative scan, the original belief is too absolute.